← Back to community benchmarks

gpt-oss-20b-MXFP4-Q8

M5 (10c) · 32 GB · 4bit · 2026-05-18

Performance

4k

tokens

965.6

PP tok/s

33.8

TG tok/s

4242

TTFT (ms)

11.8

Peak mem (GB)

Hardware

Chip M5 (10c)

Memory 32 GB

GPU Cores 10

Software

oMLX v0.3.9.dev2

macOS macOS 26.4.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	709.8	35.0	11.7 GB	view
4k	965.6	33.8	11.8 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	35.0	1.00×
2×	49.5	1.41×
4×	58.0	1.66×